Vague Terminology Linked To Poor Science

By Neuroskeptic | July 11, 2013 3:25 pm

Everyone knows that ‘correlation is not causation’ – that just because two things tend to happen together, it doesn’t mean that one of them causes the other.

However, what few people realize – except scientists – is that there’s a handy exception to this rule. This is when the two things are ‘linked’.

You see, if two things are correlated, then they can be described as ‘linked’. But if one thing causes the other, that also makes them ‘linked’.

So correlation = linked = causation, or in other words: correlation = causation! The logic’s unassailable, I’m sure you’ll agree.

*

‘Linked’ is a very popular word in neuroscience and biology. A great many things have been linked to each other and the word often serves, I fear, to make science less scientific.

If the evidence doesn’t support a certain causal association, but you want to imply that it does, you can use ‘linked’. If all the signs do point to causation, and you want to downgrade it to a correlation, you can use the L-word. It’s very versatile.

Correlation and causation are quite distinct. In describing them we ought to be clear. A wide gulf seperates them, but ‘linked’ bridges the gap: it’s a vague term that, in failing to specify the nature of an association, allows it to be read as anything.

‘Linked’ is not alone however. A linked phrase is ‘involved in’, as in “Brain region X is involved in memory”.

This is not quite as bad, as it does imply some kind of causation, but it’s often used in such a nonspecific way that it’s all but meaningless.

For instance:

  • Hitler and Churchill were both involved in WW2.
  • Democrats and Republicans were both involved in the election of Barack Obama.
  • Neuroskeptic and his laptop are both involved in writing this post.

All true, and all pretty pointless.

Linked and similar terms are a verbal fudge, or just linguistic gum to connect two ideas in the hopes that the reader will mentally confuse them. This kind of loose attitude to relating concepts is the stuff of poetry and rhetoric. It does science no good.

CATEGORIZED UNDER: media, science, select, Top Posts, Uncategorized, woo
  • Nitric-X

    May I add the equally handsome “X has a role in Y” phrase, which is linked to the above article and is involved in many discussion sections :D

    • Richard

      YES TOTALLY or “is involved in” (I am guilty of this, probably)

  • cusinemakaty

    мy coυѕιɴ ιѕ мαĸιɴɢ $51/нoυr oɴlιɴe. υɴeмployed ғor α coυple oғ yeαrѕ αɴd prevιoυѕ yeαr ѕнe ɢoт α $1З619cнecĸ wιтн oɴlιɴe joв ғor α coυple oғ dαyѕ. ѕee мore αт…­ ­ViewMore——————————————&#46qr&#46net/kkEj

    May I add the equally handsome “X
    has a role in Y” phrase, which is linked to the above article and is
    involved in many discussion sections :D

  • Rebecca McMillan

    The phrase “is correlated with” gets tedious when used repeatedly in a single paper, article or blog post. If I may match your scientific precision with verbal precision, “linked” indicates a connection between two or more objects, concepts or phenomena. While such connection may suggest covariance, it is mute regarding the direction of causality. Used in this this limited sense, “linked” can serve as a reasonable synonym for the more cumbersome phrase “is correlated with.”

    Given that you’ve rejected the use of “linked,” what alternative terms or phrases do you recommend for writers who wish to indicate that correlation is established but the direction of causation is not yet known?

    • http://blogs.discovermagazine.com Wilson Cheung

      That’s why most academic articles are really boring! We have to use the words with well-defined meanings!!!

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Well, I’d say that when the context makes it clear that “linked” is being used only in the sense of correlation (or as causation), then it’s OK as a way to avoid overusing the same word. The trouble is that “linked” can mean a great many things – sometimes the meaning is clear but often not. And I think it’s often used for that very reason.

  • Pingback: Vague Terminology Linked To Poor Science - Neur...

  • Wouter

    Well, we can always convert to predicate logic. That should solve this issue more or less. And give Frege a bit more credit for pointing out this problem over a century ago.

  • Bernard Carroll

    Another lazy doozy is the expression ‘dysregulated’ – as in
    the frontal-limbic network is dysregulated in depression or schizophrenia or
    take-your-pick. It is one of those gauzy terms that say everything and nothing.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Oh yes. That’s a whole other can of worms. ‘disordered’ ‘dysregulated’ ‘disrupted’ ‘deficit’… all just synonyms for ‘dodgy’.

  • jvkohl

    Sensory input from the environment is linked to adaptive evolution via de novo creation of olfactory receptor genes, which is HOW olfactory/pheromonal input CAUSES adaptive evolution. Food odors enable selection for nutrients that metabolize to pheromones that control reproduction.

    For comparison, mutation-driven evolution does not represent cause and effect. Natural selection is for food and sexual selection is for the pheromones produced by a potential mate — not for mutations.

  • http://neuroautomaton.com/ Zachary Stansfield

    Unfortunately, the discussion of non-causal relationships as “linked” has become the standard terminology in scientific writing. If I want to propose a novel and controversial hypothesis, all I have to do is quote a recent correlational study as “linking” or “associating” X to Y and, suddenly, my conjecture is endowed with new significance!

    I wish I could still count the number of papers I have read that report an explicitly non-causal “association”, only to go on at length in the discussion section as though causality had indeed been demonstrated. Authors can get away with these verbal hi-jinks by simply making token references to “limitations” or “future research”, which will purportedly uncover the basis of this link. Meanwhile, the limitations are never overcome and the follow-up research simply ends up adding to the “growing evidence of [non-causal] links between X and Y”.

    Just as there has been a proliferation of bad science–driven by the necessity that thousands of researchers must publish SOMETHING–so too has there been a proliferation of ambiguous language to hide the shameful lack of value in many of these efforts.

    Here’s my neologism for the day: “linkage inflation”, when correlations are promoted to causal status via mentally dulling verbiage.

  • Bradamant

    I was grinning madly as I read along until I got to the sentence: “This kind of loose attitude to relating concepts is the stuff of poetry and rhetoric.” Whoops! The author betrays both an unworthy bias and a sad (though terribly amusing) lack of knowledge about what makes for good poetry and -much more important – what rhetoric IS. Oh, the delicious irony.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Mmm. Show me a poem where the relationships between the concepts and images are completely clear. It wouldn’t be poetry. It would have none of the depth and magic that poetry requires.

      I’m not saying perfect scientific clarity is how all writing should be. That would be appalling. But it’s how science should be.

      • Zonodon

        I think that’s quite true (I know this was two weeks ago but I think this is really important). It’s totally worthwhile to defend art against people who don’t understand it but think it’s inferior to science. I don’t think the article really pushes that.

        Poetry works through implication. The greatest poems point unequivocally towards something the author observed without saying what it is outright, which allows you to feel what’s being pointed towards viscerally instead of initially connecting to it intellectually. You could work with the poetic image and how it reads to you and come to an intellectual understanding of it, but it hits you first on a visceral level.

        Science goes the other way ’round–it describes something observed as explicitly as possible, in an ideal world leaving nothing to subjective interpretation on the part of the reader. It starts in the intellect. Like a poem, you can move away from where you started, in this case towards having an emotional reaction–but that emotional reaction to the research can only fully emerge after you’ve grasped the research intellectually.

        So, raw explicitness is something to avoid in poetic writing but strive for in scientific writing. I think the only quibble I have is that the best poems are just as precise as the best scientific writing, just implicitly so instead of explicitly. Even if their meanings are not stated outright they should be totally unavoidable for the reader. Vagueness is deadly in both art and science.

  • http://chrisfilo.tumblr.com/ Chris Filo Gorgolewski

    Seems relevant here: http://xkcd.com/552/

  • Sean Lamb

    “So correlation = linked = causation, or in other words: correlation = causation! The logic’s unassailable, I’m sure you’ll agree.”

    This is a silly non issue. Linked does not imply causation, at worst it is slightly waffly way to describe correlation, although you could argue it describes a situation where you are certain that a correlation is not a false positive.

    There are plenty of situations where if A and B are genuinely correlated it is too simplistic to say A causes B or B causes A but they may interact in a more complex fashion and hence be linked, ie you know there are interactions between A and B but the exact nature of these interactions has yet to be determined.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      OK. Would you rather take a drug whose use had been correlated with cancer, or a drug linked to cancer?

  • Pingback: Vague Terminology Linked To Poor Science - Neur...

  • Mike

    This is interesting when compared with my field, geology, where we can rarely categorically show causation (due to long-lived events now being past history), but we can amass evidence to support it. So using ‘linked’ in geology is deliberately to imply causation really. But in actual fact, we are probably dealing with ‘shades of correlation’ – it’s certainly a fraught aspect of scientific philosophy in geology, one that I imagine many geos prefer not to think about!

  • Pingback: Vague Terminology Linked To Poor Science – Neuroskeptic | DiscoverMagazine.com | lit! * (Commerce & Arts)

  • Pingback: The effects of vague terminology in neuroscience | Psychology and Neuroscience Research

  • Pascal Wallisch

    You might have won the linked battle, but it looks like the linked war is a lost cause: http://books.google.com/ngrams/graph?content=linked&year_start=1800&year_end=2000&corpus=15&smoothing=3&share=

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Neuroskeptic

No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.

ADVERTISEMENT

See More

@Neuro_Skeptic on Twitter

ADVERTISEMENT
Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »